As an animal behaviorist, science and animal behavior are always on my mind; which is why I came up with the idea for this blog in the middle of my deep-water running workout yesterday. I was listening to a podcast of NPR’s Fresh Air, while speeding along at an incredibly slow pace when an interesting question was posed by the guest, neurologist David Linden, author of The Compass of Pleasure: How Our Brains Make Fatty Foods, Orgasm, Exercise, Marijuana, Generosity, Vodka, Learning, and Gambling Feel So Good.
Why are Cigarettes More Addicting Than Heroin?
He asked, “Why is it that 80% percent of all people who try cigarettes become addicted whereas only 35% of heroin users become addicts?”
This seems counter-intuitive since heroin provides a potent euphoric rush within 15 seconds of injection whereas the pleasurable feeling one gets from cigarettes is comparatively tempered. Both items trigger the pleasure circuitry of the brain and the release of the “pleasure” neurotransmitter, dopamine.
One might guess that the causes are social. That because heroin is illegal and can have more serious consequences, people are better at inhibiting their desire.
But according to Linden, there’s a neuropharmacological reason.
“Addiction is a form of learning,” he explained. “Imagine by analogy that we have a dog and we want to train him to come when called. We call and when he comes we give him a 10 oz steak we had behind our back. He eats it and then it’s gone for 12 hours. That’s similar to heroin. [Heroin] is a very large reward but even a stone junky is unlikely to shoot up more than 3 times a day. So it’s very intermittent.”
He contrasts this to the cigarette smoker who smokes a pack a day. At 20 cigarettes per pack and 10 puffs per cigarette, that’s like cutting the steak into 200 pieces so that you have 200 small rewards rather than one big one. Now when the dog comes when called he gets a tiny morsel of steak but then you practice repeatedly for reliable rewards.
The result? Says Linden, “Well by the end of the day the dog will have learned to come when called.” Similarly with cigarettes he states, “we have extremely reliable small rewards over and over that are associated with the act of puffing the cigarette and so we are extraordinarily good trainers of our inner dog.”
How To Apply this Concept Universally to Dog Training
And there we have it. A scientific analogy as to why high rates of reinforcement are important—they lead to repeated stimulation of the brain’s pleasure circuits leading to a strong desire to want to repeat the rewarded behaviors. So how does this affect how one might train a dog, or a horse, or even a giraffe? It means that we should practice and reward goal behaviors frequently so the reinforcement rate is high.
But what does this actually look like? Well, the way it looks to me may be a little different from how it looks to you or even to general population of dog trainers. So here’s my version of that high reinforcement rate.
Method 1: Sequential rewards.
Sequential rewards can provide a quick and dirty method for getting desired behavior to stick… and fast! I first started using sequential rewards when working on a research project to test a protocol for training dogs to run to a rug and remain lying calmly even with high distractions —toys, people running around or playing, or people coming to the door. This was when I was developing the Treat&Train® (now the MannersMinder®) for the Sharper Image. I worked through a number of different protocol variations but found that what worked the fastest and most reliably for training dogs to lie down and remain stationary both with and without distractions was to give sequential treats to them while they were lying down. Treats would first come at a fast rate—3 second intervals—but as soon as the dogs could reliably remain lying down with treats coming this fast we increased the interval between treats to 5, 7, 10, 15 seconds and so on. And as a result of the high reinforcement rates, dogs with no experience at lying down on cue could learn a down-stay within a few days. They purposely sought out the rug and the remote-controlled treat dispensing because they had had so many pleasurable experiences on the rug and near the machine when they were lying down.
Since then I’ve applied the sequential rewards technique liberally to other exercises especially when the dog is first learning. For instance, when I first teach dogs to say please by sitting automatically, as soon as the dog sits I provide a sequence of rewards—the first for sitting and the rest for remaining seated. Dogs generally figure out within minutes that sitting is really fun. Then we can extend this to other “say please by sitting” exercises such as, sitting to go out the door, to get a leash on, to have access to you, for the opportunity to pick up a dropped item from the floor. As a result, dogs can learn these behaviors in record time.
Similarly with come when called, although I may give a huge reward occasionally, generally I provide a series of small rewards —the first for coming and the rest for staying focused on me.
This technique of using sequential rewards also works nicely across species. For instance when training giraffes to step on a board so that their hooves could be trimmed, a group of summer interns that I supervised first gave a series of treats every 2-3 seconds once the giraffe had one foot on the board. When the giraffes reliably kept their hoof on the board with treats coming this frequently, we increased the interval between treats so they had to stay on the board longer and longer to earn the same amount of reward.
Method 2: Frequency
In addition to giving a string or series of small treats when the animal is first learning certain behaviors, it’s also important to practice the behaviors frequently. Linden’s analogy of 1 big vs. 200 small treats really hits home because it’s one of the keys principles I use for really fast training. I recommend owners use all of their dog’s regular kibble and reward the dog kibble by kibble when they are home interacting with the dog instead of with an entire bowl all at once.
Specifically I recommend owners reward their dogs for automatically sitting throughout the day because if the dogs are rewarded 200 times for this one behavior then within a day or two, their sitting behavior should be automatic and strong. This is backbone of my version of the Learn to Earn Program and it provides a quick way to train dogs to look to your for guidance and to gain self-control.
How will you use the story of heroin and cigarettes to improve your training?
There are probably many variations on how you can use the information, but I think the take-home message is clear. Some people may feel that rewards are a somewhat weak way to train dogs, and even people. However, when the right rewards are used consistently and predictably at a high reinforcement rate, they can become extremely strong, even addicting.